Many of the investigations in this genetic epidemiology project arise from observations in families at high risk of cancer or in other etiologic studies. Case-control study data from 718 non-Hispanic white patients with invasive cutaneous melanoma from melanoma clinics in Philadelphia and San Francisco and 945 matched controls from outpatient clinics with similar catchment areas to the cases are being used to develop a risk assessment model for estimating absolute risk of melanoma. Previously, DNA repair capacity (DRC) in lymphocytes from subjects from a case-control study of 183 incident melanoma cases and 179 controls conducted in North-Eastern Italy showed that DRC did not differ between cases and controls overall, but subjects with low DRC and poor tanning ability or dysplastic nevi were at increased risk of melanoma. To explore the genetic basis of these findings, a series of SNPs in DNA repair genes were analyzed. SNPs found in DNA polymerase or abasic repair genes were associated with reduced melanoma risk, while SNPs in NER genes were associated with higher risk in older subjects, or with decreased DRC. Standardized instruments were used to measure subjects constitutive skin color and UV sensitivity, and the role of SNPs was analyzed in potential pigmentation genes. The results showed that both colorimeter-based skin brightness and variants of the MC1R pigmentation gene were associated with melanoma risk, and the risk increased with the number of MC1R variants, even after adjustment for the traditionally assessed pigmentation characteristics. MC1R variants were also associated with melanoma thickness in the cases. As a follow-up to a DCEG comprehensive case-control study of adults with brain tumors, a family-based study of the parents, siblings and adult children of the 480 eligible glioma cases has been conducted. Relatives of 365 of the glioma cases were interviewed about personal and family medical history and other risk factors and were asked to provide buccal cells as a source of DNA. Analyses to examine the risk of cancer among the first degree relatives compared to population controls are in process. In the future, the relatives will be used as controls for the glioma cases in association studies and in analyses to evaluate the roles of genetic susceptibility and environmental exposures on the risk of gliomas and etiologically related tumors. Retrospective review of clinical data comparing 33 medulloblastoma patients from a single institution to their 46 unaffected relatives revealed a paucity of clinical findings among the majority of medulloblastoma patients. The results suggest that clinically recognizable syndromes are uncommon among patients with medulloblastoma, however, PTCH1 and SUFU mutations were present at a low but significant frequency.Registry data from Sweden and Denmark are continuing to be analyzed. A paper in press describes the risk of lymphoproliferative (LP) tumors in relatives of patients with non-Hodgkin lymphoma (NHL). Relatives were at significantly increased risk for NHL (hazard ratio=1.7, 95% CI, 1.4-2.2) and Hodgkin lymphoma (HL, hazard ratio=1.4, 95% CI,1.0-2.0). The risk for chronic lymphocytic leukemia (CLL) was increased but was not significant. There was no increased risk of multiple myeloma (MM). Age at diagnosis of case proband did not affect risk in relatives. Relatives of cases with an aggressive subtype of NHL were at further increased risk for developing NHL (hazard ratio=3.6, 95% CI, 1.8-7.0). A manuscript describing the results for MM cases, controls and relatives has been submitted. Relatives of MM cases are at significantly increased risk for developing MM (hazard ratio=1.7, 95% CI, 1.0-2.7) but not for other LP tumors. The association of autoimmune diseases and LP tumors in cases, controls, and relatives has also been examined.